Our survey aimed to investigate avian influenza (AI) and Newcastle disease (ND) prevalence and risk factors in three areas of Mali at risk for occurrence of H5N1 highly pathogenic avian influenza. Blood samples and cloacal and oropharyngeal swabs were collected from 1470 birds between February 2007 and May 2008 and were tested by commercial enzyme-linked immunosorbent assay to detect antibodies and real-time reverse-transcription (rRT)-PCR to detect virus. Risk factors associated with seropositivity or positive rRT-PCR were identified by random effect logistic regression. AI seroprevalence was significantly lower in birds from commercial farms (0%) than in village backyard birds (3.1%). For backyard birds, no individual risk factors (species, age, sex) were identified, but birds in the Mopti area in the Sahelian zone, where millions of wild birds migrate, were more seropositive than in the Sikasso area in the Sudano-Guinean zone (odds ratio [OR] ?=? 2.0, P ?=? 0.051). Among backyard birds nonvaccinated against ND, ND seroprevalence was 58.4%, and the odds of seropositivity was 2.0 higher in chickens than in ducks, 1.7 higher in females than in males, 3.1 higher in adults than in young birds, and 3.0 higher in poultry from the Sikasso area than from the Mopti area (P < 0.01 in all cases). Prevalence established by rRT-PCR was low for both AI virus (1.1%) and ND virus (2.6%) and was associated with no risk factors for AI but was higher in chickens than in ducks (OR ?=? 5.3, P ?=? 0.05) and in the Sikasso area than in the Mopti area (OR ?=? 3.4, P ?=? 0.027) for ND. For AI and ND, prevalence assessed by serology or rRT-PCR varied over time, although seasonal and interannual variation could not be clearly distinguished. The intracluster correlation coefficient for serologic data was low for AI (0.014) and higher for ND (0.222). These results are useful to optimize surveillance and control strategy for notifiable avian diseases in African countries with similar agroec